It is claimed that an uncountable product of $[0,1]$ is not sequentially compact, e.g. in Wikipedia (I think replacing $[0,1]$ by $\{0,1\}$ doesn't make much difference). However, the constructions I saw always take an uncountable set of the same cardinality as the reals. So I wonder whether $\{0,1\}^{\omega_1}$ is also non-sequentially compact, and if yes, whether one can construct explicitly a sequence without convergent subsequence. Constructing such a sequence in $\{0,1\}^\mathfrak{c}$ requires no choice; does one need some form of choice to say something about $\{0,1\}^{\omega_1}$?


A family $\mathscr{S}\subseteq[\omega]^\omega$ is a splitting family if for each infinite $A\subseteq\omega$ there is an $S\in\mathscr{S}$ such that $A\cap S$ and $A\setminus S$ are both infinite. The splitting number $\mathfrak{s}$ is defined to be the smallest cardinality of a splitting family. It’s easy to prove that $\omega_1\le\mathfrak{s}\le 2^\omega=\mathfrak{c}$. Theorem $6.1$ of Eric K. van Douwen, ‘The Integers and Topology’, in The Handbook of Set-Theoretic Topology, K. Kunen and J.E. Vaughan, eds., North Holland, $1985$, says that

$$\begin{align*} \mathfrak{s}&=\min\{\kappa:\{0,1\}^\kappa\text{ is not sequentially compact}\}\\ &=\min\{\kappa:\text{there is a compact space of weight }\kappa\text{ that is not sequentially compact}\}\\ &=\min\{\kappa:\text{there is a ctbly cpt space of weight }\kappa\text{ that is not sequentially compact}\}\;. \end{align*}$$

If $\kappa$ and $\lambda$ are regular cardinals with $\omega_1\le\kappa\le\lambda$ it is consistent with $\mathsf{ZFC}$ that $\mathfrak{s}=\kappa$ and $2^\omega=\lambda$; this is (part of) Theorem $5.1$ of the same monograph.

In short, it $\{0,1\}^{\mathfrak{c}}$ is never sequentially compact, but $\{0,1\}^{\omega_1}$ can be, consistently.

  • $\begingroup$ Also, the Wikipedia page for sequential compactness is misleading: it says "the product of uncountably many copies of the closed unit interval" is not sequentially compact; by the result you quote in your answer (especially the second line), it is consistent with ZFC that $[0,1]^{\omega_1}$ is sequentially compact (the Wiki page cites the book of Steen and Seebach, which in fact discusses $\mathfrak{c}$ many copies of $[0,1]$ - not merely $\omega_1$). This (or an equivalent form thereof) has driven me nuts on and off for quite some time - thank you so much for this answer! $\endgroup$ – Philip Brooker Nov 7 '13 at 1:19
  • $\begingroup$ @Philip: You’re very welcome. And I just now corrected that Wikipedia entry. $\endgroup$ – Brian M. Scott Nov 7 '13 at 9:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.